Background
Glycated haemoglobin A1c (HbA1c) is used in clinical medicine as an indicator of average blood glucose levels over three months, in addition to its proposal as a diagnostic and screening tool for diabetes. However, its role in predicting cardiovascular outcomes in the general population remains uncertain. 

Purpose
Diabetes mellitus is regarded as a classical risk factor for cardiovascular disease. Due to the clinical need to identify novel risk factors to improve cardiovascular risk prediction HbA1c may be a potential candidate.

Methods
Data from six prospective population-based cohort studies across Europe comprising 36180 participants were analysed. HbA1c was evaluated in conjunction with classical cardiovascular risk factors (CVRFs) for association with cardiovascular mortality, cardiovascular disease (CVD), and overall mortality in the cohort of participants, as well as in non-diabetic (N=32477), and diabetic subjects (N=3703). 

Results
Kaplan-Meier curves showed a higher event rate with increasing HbA1c levels. Cox regression analysis revealed a significant association of continuous log-transformed HbA1c levels per one unit increase, with a hazard ratio (HR) of 1.79 (95% confidence interval (CI) 1.27−2.51,p<0.001) for cardiovascular mortality,1.62 (95% CI 1.25−2.09,p<0.001) for CVD, and 1.56 (95% CI 1.29−1.90,p<0.001) for overall mortality. The time-to-event analysis revealed a predominantly linear association (Figure 1). An increased risk of CVD was observed in subjects without diabetes with increased continuous log-transformed HbA1c levels per unit (HR 1.53; 95% CI 1.10,2.13;p=0.0012).
An HbA1c cut-off value of 39.9 mmol/mol (5.8%), 36.6 mmol/mol (5.5%), and 38.8 mmol/mol (5.7%) for cardiovascular mortality, CVD, and overall mortality, respectively, was determined for selecting individuals at an increased risk.

Conclusions
HbA1c was demonstrated to be an independent prognostic biomarker for all investigated outcomes in the general European population. A linear relationship was observed between outcomes and an increase of HbA1c levels. Subjects with an HbA1c level below the threshold for diagnosing diabetes had an increased risk, underlining the importance of prediabetes. 

Figure 1 Penalised cubic splines for the association between HbA1c and time-to-event. Untransformed HbA1c in mmol/mol are shown on the x-axis in parentheses. The natural logarithm was used. Importantly, for the correct interpretation of the depicted graph, only 1% of individuals had HbA1c levels <4.0% (20 mmol/mol) and 0.3% with HbA1c levels >10.3% (90 mmol/mol).